bug#6232: Suggestion for isearch: in backward search yank to front of search string

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Lennart Borgman]

Yanking to the end of the search string in isearh (i.e. C-w etc.) is
very practical. However I miss yanking to the front of the search
string.

My suggestion is to change the semantics a bit and yank to the front
of the search string when searching backwards. What do you think?

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Juri Linkov]

> Yanking to the end of the search string in isearh (i.e. C-w etc.) is
> very practical. However I miss yanking to the front of the search
> string.
>
> My suggestion is to change the semantics a bit and yank to the front
> of the search string when searching backwards. What do you think?

I think this would be wrong to do because C-w and all other yanking commands
always yank in the forward direction.

-- 
Juri Linkov
http://www.jurta.org/emacs/

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Lennart Borgman]

On Fri, May 21, 2010 at 2:09 AM, Juri Linkov <juri@jurta.org> wrote:
>> Yanking to the end of the search string in isearh (i.e. C-w etc.) is
>> very practical. However I miss yanking to the front of the search
>> string.
>>
>> My suggestion is to change the semantics a bit and yank to the front
>> of the search string when searching backwards. What do you think?
>
> I think this would be wrong to do because C-w and all other yanking commands
> always yank in the forward direction.

My suggestion to use the normal key bindings but let them act in the
reverse direction is based on reasoning like this:

1) There is a lack of key bindings in isearch-mode.

2) Using the same keybindings make them easy to remember.

3) It does seem rather logical that the yanking is done at the other
end when search direction is reversed.

4) Maybe more important: The marker in isearch that shows the
direction will be at the position where the yanking is done. Or at
least nearly. So the semantics is in a sense the same since yanking is
done forward from that marker.

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Andreas Schwab]

Lennart Borgman <lennart.borgman@gmail.com> writes:

> 2) Using the same keybindings make them easy to remember.
>
> 3) It does seem rather logical that the yanking is done at the other
> end when search direction is reversed.

It would be totally confusing behaviour.  Normal characters are always
added at the end, so yanking should always behave the same.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Lennart Borgman]

On Fri, May 21, 2010 at 10:34 AM, Andreas Schwab <schwab@linux-m68k.org> wrote:
> Lennart Borgman <lennart.borgman@gmail.com> writes:
>
>> 2) Using the same keybindings make them easy to remember.
>>
>> 3) It does seem rather logical that the yanking is done at the other
>> end when search direction is reversed.
>
> It would be totally confusing behaviour.  Normal characters are always
> added at the end, so yanking should always behave the same.


I can see your point, but my gut feeling is perhaps not the same.

Also consider the 4th point I made:


> 4) Maybe more important: The marker in isearch that shows the
> direction will be at the position where the yanking is done. Or at
> least nearly. So the semantics is in a sense the same since yanking is
> done forward from that marker.


I any case I think it is important that the key sequences for adding
at the beginning and end of the search string are similar. There are 4
yank commands in isearch for adding at the end. How about adding shift
to these to add at the beginning?

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Glenn Morris]

There was fairly strong opposition to this, so closing as "wontfix".

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Richard Stallman]

In backward search, self-inserting characters add to the end of the
search string.  To make yank add at the beginning would be
inconsistent and not useful.

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Lennart Borgman]

On Fri, May 21, 2010 at 11:09 PM, Richard Stallman <rms@gnu.org> wrote:
> In backward search, self-inserting characters add to the end of the
> search string.  To make yank add at the beginning would be
> inconsistent


Both you and Andreas think that yanking at the beginning for backward
isearch is inconsistent.

If you both feel that way I tend to agree (since this is a usability problem).


> and not useful.


However if it is useful is a quite different thing. I have use cases
for it. I can't imagine that other people does not have it so I assume
this is not really what you (or Andreas) wanted to say.

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Andreas Schwab]

Lennart Borgman <lennart.borgman@gmail.com> writes:

> Yanking to the end of the search string in isearh (i.e. C-w etc.) is
> very practical. However I miss yanking to the front of the search
> string.
>
> My suggestion is to change the semantics a bit and yank to the front
> of the search string when searching backwards. What do you think?

Such a feature should definitely be on a separate key.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

bug#6232: Suggestion for isearch: in backward search yank to front of search string

[Leo]

On 2010-05-21 09:27 +0100, Andreas Schwab wrote:
> Lennart Borgman <lennart.borgman@gmail.com> writes:
>
>> Yanking to the end of the search string in isearh (i.e. C-w etc.) is
>> very practical. However I miss yanking to the front of the search
>> string.
>>
>> My suggestion is to change the semantics a bit and yank to the front
>> of the search string when searching backwards. What do you think?
>
> Such a feature should definitely be on a separate key.
>
> Andreas.

C-y is almost useless to me so I have done something like this:

(define-key isearch-mode-map [remap isearch-yank-line] 'isearch-yank-sexp)

(defun isearch-yank-sexp ()
  "Pull the sexp at point into search string."
  (interactive)
  (if (and isearch-just-started (bounds-of-thing-at-point 'sexp))
      (progn (goto-char (car (bounds-of-thing-at-point 'sexp)))
             (isearch-yank-string (thing-at-point 'sexp)))
    (isearch-yank-internal
     (lambda () (ignore-errors (forward-sexp 1)) (point)))))

Leo